SEM: p-value of direct, indirect and total effects (putexcel)

[Nick Bornschein]

Hi,

I'm calculation a SEM with variable x->y, x->z and z->y. Using "estat teffects, standardized", I get the effects I want to look at in a table in Stata result window, including the p-values of these coefficients. I wanted to use putexcel to export my results but looking with "(e)return list", I can't find any matrice which contains the p-values of the teffects-table. Are these results not stored?

-Nick

[Clyde Schechter]

The p-values shown in -estat teffects- are not stored. What you can do is grab the matrices that are returned in r(), which give, collectively, the coefficients and covariances. Then you divide each coefficient by the square root of its variance to get a z-statistic, and then calculate a two-tailed p-value as 2*normal(-abs(z_statistic)). You can then pass those results to -putexcel-. Yes, it's a lot of work!